Fault Tolerance by Quartile and Outlier Methods in Wireless Sensor and Actor Networks
Autor: | Wei-Shiou Chang, 張瑋修 |
---|---|
Rok vydání: | 2009 |
Druh dokumentu: | 學位論文 ; thesis |
Popis: | 97 Wireless Sensor and Actor Networks (WSAN) is a new wireless heterogeneous sensor network, composed of sensor nodes and sink and few of actors. Actors were equipped with more battery power, better processing capabilities, and larger transmission radius. Actors can reduce the processing time of events and enlarge the sensing area in WSAN. Since nodes are prone to failure due to energy depletion, and communication link interfered, designing an efficient fault tolerance mechanism becomes an important issue in WSAN. Most researches focused on link fault tolerance without considering sensing data fault tolerance. However, actor may take an incorrect action when receiving error data, delivered by sensor nodes with invalid sensing capability. To solve the problem, we proposed a fault tolerance by quartile and outlier methods (FTQOM). In WSAN, FTQOM first judged the range of correct data by quartile method, and actors responded correct actions according to the correct sensed data. However, quartile method cannot exclude and receive error sensed data sent by in valid nodes at next time. To resolve this problem, outlier method was invoked to detect and exclude the invalid sensor nodes. Hence, actors can save the processing time by directly excluding the error sensed data at next time. Moreover, the fault tolerance mechanism may be more complete by integrating with link fault tolerance scheme proposed by other researchers. We define three metrics: detected correct data rate ( ), detected data accuracy ( ), and detected error value ratio ( ) to measure and compare the performance of FTQOM and AFTM (Average Fault Tolerant Mechanism). From the simulation results, FTQM was better than AFTM by 55.8% on the, by 81.1% on the, and by 125.4% on the, respectively. FTQOM was better than FTQM by 0.3% on the, by 0.3% on the, by 0.5% on the, respectively. Under the modulating the value of tolerance, FTQM was still better than AFTM by 84.0% on the, by 56.6% on the, and by 1.7% on the, respectively. FTQOM was better than FTQM by 0.4% on the, by 0.3% on the, and by 1.6% on the, respectively. |
Databáze: | Networked Digital Library of Theses & Dissertations |
Externí odkaz: |